Two of the new file-organization concepts that Windows Vista has brought to the table are outgrowths of existing ideas that may be familiar to some people but not to others: tagging and stacking (or stacks). In this edition of the XP Guide to Vista I'll be exploring these two features in tandem.
A review of Office 2007, which I worked on along with several other authors, is up on TechWeb. (I wrote the Word, Outlook and Access portions of the review.)
See what you think!
Over at Ed Bott's weblog there's a post about a Vista issue that I was going to discuss at some point: Vista's defragger has no GUI. It has a scheduler (it runs automatically once a week) and a manual start, but no progress meter or defrag display. To wit:
At first I resisted this with a great deal of annoyance: what's the point of having a tool like this if we can't even see the progress meter?
Here is why this issue is not as obvious as it might seem.
The more I learn about fragmentation on a drive, the more I realize there is no one convenient metric that you can use to describe how badly a drive is fragmented. You could talk about individual file fragments, but what if they're for files that are rarely accessed? Or free space fragmentation, but what if the free space "fragments" are so large (50MB or more) that it's not going to create an issue for any new file written in it? I guess it might be possible to come up with some kind of global "fragmentation index" -- like the way many of the newer anti-malware programs have some kind of "threat index" or "system protection index", but in the end I'm not sure that gives us any more useful information than we had before. If a drive is fragmented badly enough that it's impacting performance, then we tend to take a fire-and-forget approach to the whole thing: we defrag the drive and move on to doing more serious work.
One of the reasons for eliminating the display comes straight from the horse's mouth:
The new interface seems “dumbed down.” Why remove all the detail?
...one of the biggest and consistent complaints we had from users (broad sample here from home users to experienced IT Pros) in the past was that a vast majority of them had no idea what the detailed fragmentation statistics they saw meant. The Windows XP graphical view also had some limitations and inaccuracies that prevented it from being included in Windows Vista. If you really want to keep a close eye on fragmentation, I’d recommend using the command-line tool Defrag.exe.Why was the defrag progress indicator removed?
Part of the problem with the Windows XP defrag tool was that percent complete was not accurate or meaningful. Depending on the phase of defrag, 1% of progress could take from several seconds to minutes, which made the progress indicator highly unreliable. The difficulty here is that since defrag is a multi-pass process (multiple iterations of file defragmentation and free space consolidation) there is no way to accurately predict when defrag will complete since the number of loop iterations and how long each takes are highly dependent on the layout of the files on the volume, the level of file and free space fragmentation, and the other system activity. While I agree that having no progress is bad, misleading progress I believe is worse. Also, the idea behind the new automated defrag is that users will not have to think about it not worry about the progress it is making. With defrag running regularly, the system will be close to optimal levels of fragmentation, and subsequent defrag runs should not take long.
To which people have responded: why not simply rework the display so it does make sense? My answer to that, in turn, is part of the above: perhaps fragmentation is simply not something you can express graphically in a convenient way without the data being misleading. This goes back into something I have encountered in other contexts -- the idea that too much of the wrong kind of information is just as bad as no information at all. One classic example of this is the Homeland Security "threat level" index, which tells you absolutely nothing you didn't already know and never drops below 2 anyway.
I found that with Vista, not having a GUI to look at in the defragger didn't make me want to manually defragment the drive as much. There's actually an odd parallel here with the Office 2007 ribbon interface: I focused less on Word itself and more on my document, which is probably the point. Sometimes more is not, in fact, better, especially if it leads into doing something counterproductive.
My point is that the less I have to think about something as nuts-and-bolts as the level of fragmentation on my drives, the better. I can still get fragmentation stats about a given drive from the command-line version of the app if I really need them, but I've never needed them in more than a cursory, informational way; it's not as if I can use them to further guide the actions of the program in some directed fashion. And even if I could, who's to say I wouldn't simply be making things worse by second-guessing?
Something I've always liked about Firefox is the community of developers who create custom builds of the program -- they take the source code and compile it with specific optimizations. Among the very best is the Stipespeed builds, created with optimizations for the most modern processors and markedly faster than the stock Firefox build right out of the box.
Stipe released his build of Firefox 2.0 back at the end of October. If you want to use it and you already have an existing Firefox install, all you need to do is unpack it into a directory by itself and run the firefox.exe executable. It'll attempt to register itself as "Bon Echo" (the development codename for 2.0; community builds of the program cannot officially be called "Firefox"). This way you can run Stipespeed in parallel with the original install; if you don't like it, you can always delete it and revert to the original by running the original application from its own directory. And since user preferences are stored in the user profile and not with the application, it's all the more convenient.
[Note: Vista users, you'll need to run Stipespeed as Administrator to get it to register as the default web browser.]
One of the things I've always found problematic about Firefox (even though I'm quite fond of it in general) is how it's not possible to launch FF windows in their own processes. All FF windows are shared by the same, single instance of the firefox.exe image, so if one crashes, they all crash.
I did some digging and found out that it is possible to launch a new instance of Firefox in a separate process, but only if it uses a different profile. The reason for this, I suspect, is because the profile is locked for exclusive use by one instance of the program and can't (yet) be shared between instances.
If you have multiple profiles, though, and you want to use them side-by side, all you need to do is create a batch file to launch FF that way, like so:
@echo off
set MOZ_NO_REMOTE=1
start firefox.exe -p
set MOZ_NO_REMOTE=0
This sets the MOZ_NO_REMOTE environment variable, which FF uses to detect if it's supposed to launch in a new process or not. When it launches, though, it'll ask you what profile you want to use; if you attempt to use the profile that's currently in use, it'll forbid it.
I should point out that if FF 2.0 crashes, it does attempt to re-open the windows you had open at the time when you next launch the program. This is at least partial compensation for not having multiple instances, but it's still not quite the same thing.
As many of us know, Windows XP had its own indexed search system, but it was one of XP's most underused and badly-implemented idea. Vista's systemwide search system takes the same basic idea -- crawl the most commonly-used directories of the user's PC for content which can be indexed for fast searching -- and gets it mostly right.
The bad news: if you're used to the way XP does searching, you're going to need to learn how to do this from scratch. The good news: the new search system, while it does have its quirks, can be learned and modified relatively easily. The best news: it works like magic.
Finally! Windows Vista has been released to manufacturing at last (with a build number of 6000.16386.061101-2205, if you're a trivia hound).
As a certain dead rock star once said, what a long, strange trip it's been. The trip to Vista has been long and difficult, and it was disheartening to see Microsoft fall back and retrench as many times as they did. A great many features that were meant to make the final cut had to be dropped, reworked, or abandoned entirely.
But what has been delivered is in many key ways a solid improvement on XP, and in a few key ways a massive improvement. It's a mixture of the incrementalism that was Windows 95-to-98, and the revolutionary changes that was 98-to-XP. I lament the loss of the object-based file system that has been one of Microsoft's goals for nigh on a decade now, but I welcome the indexed search (which actually works and isn't just something hastily wedged in sideways), the heightened security (although I doubt we'll see the real benefits of that until Vista is the prevalent Windows OS), and many other things, big and small, that make Vista what it is.
I'm using RC1 on my notebook and plan to upgrade it via a clean install when the gold code drops into my lap. From there I'll eventually -- eventually -- upgrade my desktop machine after I make a number of checks to insure certain programs still work properly. It's done, but I'm in no hurry to upgrade completely; XP isn't going anywhere, and probably won't be for the next few years. That and I'm betting most of the upgrades to Vista will be passive -- i.e., people buying PCs pre-loaded with Windows, which is the way it's been for quite some time.
I have to be realistic. If Microsoft had tried to deliver Vista with all the features that they originally promised, with the problems they experienced, we might not even have seen a working Beta 1 yet. I don't think they realized just how much they were trying to bite off until it was too late, and they had to scale back before committing themselves to delivering something that might simply have been unusable.
Congratulations, everyone. Now take some time off. You've earned it!
A post in the Windows Vista Team Blog by Nick White seems to finally have lifted away the clouds of confusion about Vista's licensing.
I’m very pleased to let you know ... that the Windows division has revised the retail license terms for Windows Vista in a significant way. Namely, the terms regarding license-to-device assignment of the retail product (including Home Basic, Home Premium, Business and Ultimate) now read as follows:
- You may uninstall the software and install it on another device for your use. You may not do so to share this license between devices.
... You who comprise the enthusiast market are vital to us for several reasons, not least of all because of the support you’ve provided us throughout the development of Windows Vista. We respect the time and expense you go to in customizing, building and rebuilding your hardware and we heard you that the previous terms were seen as an impediment to that -- it’s for that reason we’ve made this change.
Good! This is far less ambiguous and confusing, and is exactly what most people wanted to hear: that they can uninstall and move Vista (at least, over-the-counter copies) as needed. I'm still trying to get hard details about the way the new activation system works, but it would make sense that they could make it that much more forgiving if it's being augmented by periodic WGA checks to see if someone's trying to cheat on the license terms. This is obviously something people (me included) are going to want to "kick the tires" of and see how it behaves. (Skeptic that I am, the way they describe it and the way it's implemented don't have to match up ... but I'm also an optimist, and I'd like to believe they do.)
My understanding is that the OEM version is sold with and tied to a specific device, and that this was the case with XP as well. My own computer, for instance, came with a copy of XP Professional SP2 that's licensed specifically for it, and I've had no trouble respecting that since it means one less thing I have to buy or manage separately. I think I mentioned before that the only times I've had real problems with PA or WGA was when I messed up something on my own, although I hope that Vista is robust enough that WGA is not going to spontaneously break or deliver false reports. (From what I have been told about the way WGA is implemented in Vista, there's a 30-day warning / grace period if WGA flags your system, during which time you can track down and correct any potential problems that might be causing WGA to balk.)
To the folks at Microsoft who made this decision: Thank you. You've made this a lot less confusing and potentially contradictory, and you've won back a lot of people who not only felt alienated and angry by this but had the potential to influence others.
